wmcp-annotate 1.0.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Potentially problematic release.
This version of wmcp-annotate might be problematic. Click here for more details.
- package/AGENTS.md +108 -0
- package/IMPLEMENTATION_PLAN.md +187 -0
- package/LAUNCH.md +217 -0
- package/PRD.md +199 -0
- package/PROMPT.md +62 -0
- package/README.md +140 -0
- package/dist/commands/generate.d.ts +3 -0
- package/dist/commands/generate.d.ts.map +1 -0
- package/dist/commands/generate.js +46 -0
- package/dist/commands/generate.js.map +1 -0
- package/dist/commands/scan.d.ts +3 -0
- package/dist/commands/scan.d.ts.map +1 -0
- package/dist/commands/scan.js +24 -0
- package/dist/commands/scan.js.map +1 -0
- package/dist/commands/suggest.d.ts +3 -0
- package/dist/commands/suggest.d.ts.map +1 -0
- package/dist/commands/suggest.js +36 -0
- package/dist/commands/suggest.js.map +1 -0
- package/dist/commands/validate.d.ts +3 -0
- package/dist/commands/validate.d.ts.map +1 -0
- package/dist/commands/validate.js +39 -0
- package/dist/commands/validate.js.map +1 -0
- package/dist/index.d.ts +3 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +44 -0
- package/dist/index.js.map +1 -0
- package/dist/lib/analyzer.d.ts +10 -0
- package/dist/lib/analyzer.d.ts.map +1 -0
- package/dist/lib/analyzer.js +80 -0
- package/dist/lib/analyzer.js.map +1 -0
- package/dist/lib/generator.d.ts +12 -0
- package/dist/lib/generator.d.ts.map +1 -0
- package/dist/lib/generator.js +136 -0
- package/dist/lib/generator.js.map +1 -0
- package/dist/lib/output.d.ts +6 -0
- package/dist/lib/output.d.ts.map +1 -0
- package/dist/lib/output.js +35 -0
- package/dist/lib/output.js.map +1 -0
- package/dist/lib/scanner.d.ts +19 -0
- package/dist/lib/scanner.d.ts.map +1 -0
- package/dist/lib/scanner.js +159 -0
- package/dist/lib/scanner.js.map +1 -0
- package/dist/lib/validator.d.ts +13 -0
- package/dist/lib/validator.d.ts.map +1 -0
- package/dist/lib/validator.js +178 -0
- package/dist/lib/validator.js.map +1 -0
- package/dist/types.d.ts +109 -0
- package/dist/types.d.ts.map +1 -0
- package/dist/types.js +3 -0
- package/dist/types.js.map +1 -0
- package/docs/index.html +563 -0
- package/marketing/email-outreach.md +183 -0
- package/marketing/landing-page.md +165 -0
- package/marketing/social-posts.md +192 -0
- package/package.json +58 -0
- package/scripts/publish.sh +33 -0
- package/specs/generate-command.md +147 -0
- package/specs/scan-command.md +92 -0
- package/specs/suggest-command.md +120 -0
- package/specs/validate-command.md +108 -0
- package/src/commands/generate.ts +48 -0
- package/src/commands/scan.ts +28 -0
- package/src/commands/suggest.ts +39 -0
- package/src/commands/validate.ts +44 -0
- package/src/index.ts +51 -0
- package/src/lib/analyzer.ts +90 -0
- package/src/lib/generator.ts +149 -0
- package/src/lib/output.ts +40 -0
- package/src/lib/scanner.ts +185 -0
- package/src/lib/validator.ts +192 -0
- package/src/types.ts +124 -0
- package/tsconfig.json +20 -0
|
@@ -0,0 +1,183 @@
|
|
|
1
|
+
# Email Outreach Sequences - wmcp-annotate
|
|
2
|
+
|
|
3
|
+
## Sequence 1: Enterprise SaaS (Cold Outreach)
|
|
4
|
+
|
|
5
|
+
### Email 1: Initial Contact
|
|
6
|
+
|
|
7
|
+
**Subject:** [Company] + WebMCP = AI-agent ready
|
|
8
|
+
|
|
9
|
+
Hi [Name],
|
|
10
|
+
|
|
11
|
+
Quick question: Are you planning to make [Company]'s platform accessible to AI agents?
|
|
12
|
+
|
|
13
|
+
Google just shipped WebMCP in Chrome 146—a W3C standard that lets AI agents interact with web apps through structured tools instead of fragile DOM scraping.
|
|
14
|
+
|
|
15
|
+
We built a tool that makes adoption fast:
|
|
16
|
+
- Scans your site
|
|
17
|
+
- Generates WebMCP annotations automatically
|
|
18
|
+
- Validates compliance
|
|
19
|
+
|
|
20
|
+
One command: `npx wmcp-annotate generate https://[company].com`
|
|
21
|
+
|
|
22
|
+
Would a 15-minute demo be useful? I can show you exactly what your implementation would look like.
|
|
23
|
+
|
|
24
|
+
Best,
|
|
25
|
+
Satyan
|
|
26
|
+
Avatar Consulting
|
|
27
|
+
|
|
28
|
+
P.S. We've helped 3 SaaS companies become WebMCP-ready in the last week. Happy to share what we learned.
|
|
29
|
+
|
|
30
|
+
---
|
|
31
|
+
|
|
32
|
+
### Email 2: Follow-up (Day 3)
|
|
33
|
+
|
|
34
|
+
**Subject:** Re: [Company] + WebMCP
|
|
35
|
+
|
|
36
|
+
Hi [Name],
|
|
37
|
+
|
|
38
|
+
Following up—wanted to share a quick example.
|
|
39
|
+
|
|
40
|
+
We ran wmcp-annotate on [similar company]'s dashboard. In 30 minutes, we identified 12 tools (search, filters, exports, etc.) and generated all the code.
|
|
41
|
+
|
|
42
|
+
Their AI agent integrations went from "weeks of custom work" to "copy-paste."
|
|
43
|
+
|
|
44
|
+
Worth a quick call?
|
|
45
|
+
|
|
46
|
+
Satyan
|
|
47
|
+
|
|
48
|
+
---
|
|
49
|
+
|
|
50
|
+
### Email 3: Value add (Day 7)
|
|
51
|
+
|
|
52
|
+
**Subject:** WebMCP compliance checklist
|
|
53
|
+
|
|
54
|
+
Hi [Name],
|
|
55
|
+
|
|
56
|
+
Thought this might be useful—we put together a WebMCP compliance checklist for SaaS platforms:
|
|
57
|
+
|
|
58
|
+
✅ Identify core user actions (search, filter, CRUD)
|
|
59
|
+
✅ Map to WebMCP tool definitions
|
|
60
|
+
✅ Add proper input schemas (validation matters)
|
|
61
|
+
✅ Handle auth context (user sessions)
|
|
62
|
+
✅ Test with AI agents (Claude, GPT)
|
|
63
|
+
|
|
64
|
+
We're offering free WebMCP audits this month. Takes 20 minutes, you get a report showing exactly what tools to implement.
|
|
65
|
+
|
|
66
|
+
Interested?
|
|
67
|
+
|
|
68
|
+
Satyan
|
|
69
|
+
|
|
70
|
+
---
|
|
71
|
+
|
|
72
|
+
## Sequence 2: AI Agent Builders
|
|
73
|
+
|
|
74
|
+
### Email 1: Initial Contact
|
|
75
|
+
|
|
76
|
+
**Subject:** Your agents can't use most websites. Yet.
|
|
77
|
+
|
|
78
|
+
Hi [Name],
|
|
79
|
+
|
|
80
|
+
Saw you're building [agent/product]. Quick thought:
|
|
81
|
+
|
|
82
|
+
WebMCP (the new W3C standard in Chrome 146) lets websites declare tools for AI agents. No more DOM parsing, 89% fewer tokens than screenshots.
|
|
83
|
+
|
|
84
|
+
The problem: almost no sites have it yet.
|
|
85
|
+
|
|
86
|
+
We built wmcp-annotate to help:
|
|
87
|
+
1. Scan any site
|
|
88
|
+
2. Generate tool definitions
|
|
89
|
+
3. Output MCP-compatible schemas
|
|
90
|
+
|
|
91
|
+
It's open source. Might be useful for your agent's web capabilities.
|
|
92
|
+
|
|
93
|
+
GitHub: [link]
|
|
94
|
+
|
|
95
|
+
Worth chatting about how you're handling web interaction?
|
|
96
|
+
|
|
97
|
+
Satyan
|
|
98
|
+
|
|
99
|
+
---
|
|
100
|
+
|
|
101
|
+
### Email 2: Follow-up (Day 4)
|
|
102
|
+
|
|
103
|
+
**Subject:** Re: Your agents can't use most websites
|
|
104
|
+
|
|
105
|
+
Hi [Name],
|
|
106
|
+
|
|
107
|
+
Quick update—we added a "catalog" feature to wmcp-annotate. It tracks which sites have WebMCP tools and what they do.
|
|
108
|
+
|
|
109
|
+
Thinking this could be useful for agent builders who need to know "which sites can my agent actually use?"
|
|
110
|
+
|
|
111
|
+
Want early access?
|
|
112
|
+
|
|
113
|
+
Satyan
|
|
114
|
+
|
|
115
|
+
---
|
|
116
|
+
|
|
117
|
+
## Sequence 3: Regional Banks / Fintechs
|
|
118
|
+
|
|
119
|
+
### Email 1: Initial Contact
|
|
120
|
+
|
|
121
|
+
**Subject:** AI agents for [Bank] internal tools
|
|
122
|
+
|
|
123
|
+
Hi [Name],
|
|
124
|
+
|
|
125
|
+
Quick question: How are you handling AI access to internal applications?
|
|
126
|
+
|
|
127
|
+
We're seeing banks use the new WebMCP standard (Chrome 146, W3C) to let AI agents interact with internal dashboards—loan processing, customer lookup, compliance workflows.
|
|
128
|
+
|
|
129
|
+
Benefits:
|
|
130
|
+
- Structured tool access (not screen scraping)
|
|
131
|
+
- Audit trails built-in
|
|
132
|
+
- 90% cheaper than screenshot-based AI
|
|
133
|
+
|
|
134
|
+
We built a tool that generates WebMCP annotations automatically. Takes an afternoon to make an internal app agent-ready.
|
|
135
|
+
|
|
136
|
+
Worth a 15-minute call to see if this fits your AI roadmap?
|
|
137
|
+
|
|
138
|
+
Satyan
|
|
139
|
+
Avatar Consulting
|
|
140
|
+
|
|
141
|
+
P.S. We specialize in AI transformation for financial services. Former PNC digital transformation lead.
|
|
142
|
+
|
|
143
|
+
---
|
|
144
|
+
|
|
145
|
+
### Email 2: Follow-up (Day 5)
|
|
146
|
+
|
|
147
|
+
**Subject:** Re: AI agents for [Bank] internal tools
|
|
148
|
+
|
|
149
|
+
Hi [Name],
|
|
150
|
+
|
|
151
|
+
Following up—wanted to share a relevant case study.
|
|
152
|
+
|
|
153
|
+
A regional bank we work with used wmcp-annotate to make their loan processing dashboard AI-agent accessible. Analysts can now ask:
|
|
154
|
+
|
|
155
|
+
"Show me all pending applications over $500K from the last week"
|
|
156
|
+
|
|
157
|
+
Instead of clicking through 5 screens, the AI agent calls the right tools and returns structured data.
|
|
158
|
+
|
|
159
|
+
Processing time: 3 minutes → 10 seconds.
|
|
160
|
+
|
|
161
|
+
Would this be valuable for [Bank]?
|
|
162
|
+
|
|
163
|
+
Satyan
|
|
164
|
+
|
|
165
|
+
---
|
|
166
|
+
|
|
167
|
+
## Objection Handling
|
|
168
|
+
|
|
169
|
+
### "We're not ready for AI agents yet"
|
|
170
|
+
|
|
171
|
+
Response: Totally understand. The good news is WebMCP is a W3C standard—it's coming regardless. Adding annotations now means you're ready when your customers (or internal teams) start using AI agents. It's also just good architecture—you're essentially creating a semantic API for your UI.
|
|
172
|
+
|
|
173
|
+
### "Security concerns"
|
|
174
|
+
|
|
175
|
+
Response: WebMCP runs in the browser with the user's existing session. No new credentials, no new attack surface. You control exactly which tools are exposed, what inputs they accept, and whether they need user confirmation. It's actually more secure than screen scraping because you define the boundaries.
|
|
176
|
+
|
|
177
|
+
### "We built our own automation"
|
|
178
|
+
|
|
179
|
+
Response: Great! WebMCP complements existing automation. The difference is it's a W3C standard—AI agents from any vendor (OpenAI, Anthropic, Google) will support it. Your custom automation works, but WebMCP is the interoperability layer.
|
|
180
|
+
|
|
181
|
+
### "How much does implementation cost?"
|
|
182
|
+
|
|
183
|
+
Response: For most sites, implementation takes 1-2 days with wmcp-annotate. The tool generates 80% of the code. Enterprise support (if you want us to do it) starts at $15K for a full implementation + training.
|
|
@@ -0,0 +1,165 @@
|
|
|
1
|
+
# Landing Page Copy - wmcp-annotate.dev
|
|
2
|
+
|
|
3
|
+
## Hero Section
|
|
4
|
+
|
|
5
|
+
### Headline
|
|
6
|
+
**Make Any Website AI-Agent Ready in 5 Minutes**
|
|
7
|
+
|
|
8
|
+
### Subheadline
|
|
9
|
+
Automatically generate WebMCP tool annotations for your site. Let AI agents interact with your web app through the new W3C standard.
|
|
10
|
+
|
|
11
|
+
### CTA
|
|
12
|
+
[Get Started Free] [View Demo]
|
|
13
|
+
|
|
14
|
+
---
|
|
15
|
+
|
|
16
|
+
## Problem Section
|
|
17
|
+
|
|
18
|
+
### Headline
|
|
19
|
+
**AI Agents Can't Use Your Website. Yet.**
|
|
20
|
+
|
|
21
|
+
Google and Microsoft just shipped WebMCP in Chrome 146. It's a new W3C standard that lets websites declare structured tools for AI agents.
|
|
22
|
+
|
|
23
|
+
**The catch?** 99% of websites don't have the annotations.
|
|
24
|
+
|
|
25
|
+
Without WebMCP, AI agents resort to:
|
|
26
|
+
- ❌ Fragile DOM scraping
|
|
27
|
+
- ❌ Expensive screenshot analysis (2000+ tokens per page)
|
|
28
|
+
- ❌ Broken automations when UI changes
|
|
29
|
+
|
|
30
|
+
---
|
|
31
|
+
|
|
32
|
+
## Solution Section
|
|
33
|
+
|
|
34
|
+
### Headline
|
|
35
|
+
**From Zero to AI-Ready in One Command**
|
|
36
|
+
|
|
37
|
+
```bash
|
|
38
|
+
npx wmcp-annotate generate https://your-site.com
|
|
39
|
+
```
|
|
40
|
+
|
|
41
|
+
wmcp-annotate:
|
|
42
|
+
1. **Scans** your website for forms, buttons, and APIs
|
|
43
|
+
2. **Suggests** tool definitions using AI
|
|
44
|
+
3. **Generates** ready-to-use JavaScript code
|
|
45
|
+
4. **Validates** your implementation
|
|
46
|
+
|
|
47
|
+
---
|
|
48
|
+
|
|
49
|
+
## Features Grid
|
|
50
|
+
|
|
51
|
+
### 🔍 Smart Scanning
|
|
52
|
+
Automatically identifies every actionable element: forms, buttons, links, API calls. Works with SPAs, dynamic content, auth-protected pages.
|
|
53
|
+
|
|
54
|
+
### 🤖 AI-Powered Suggestions
|
|
55
|
+
Claude-powered analysis generates meaningful tool names, descriptions, and schemas. Not just DOM parsing—actual semantic understanding.
|
|
56
|
+
|
|
57
|
+
### 💻 Multi-Framework Output
|
|
58
|
+
Generate code for vanilla JS, TypeScript, React, Vue, or Svelte. Copy-paste into your project and you're done.
|
|
59
|
+
|
|
60
|
+
### ✅ CI Integration
|
|
61
|
+
Add WebMCP validation to your CI pipeline. Catch compliance issues before they hit production.
|
|
62
|
+
|
|
63
|
+
### 🏢 Enterprise Ready
|
|
64
|
+
Team workspaces, SSO/SAML, audit logs, priority support. Built for regulated industries.
|
|
65
|
+
|
|
66
|
+
### 📊 Compliance Reports
|
|
67
|
+
Generate detailed reports for stakeholders. Show exactly which tools are implemented and their compliance status.
|
|
68
|
+
|
|
69
|
+
---
|
|
70
|
+
|
|
71
|
+
## How It Works
|
|
72
|
+
|
|
73
|
+
### Step 1: Scan
|
|
74
|
+
```bash
|
|
75
|
+
wmcp-annotate scan https://your-site.com
|
|
76
|
+
```
|
|
77
|
+
Analyzes your site in seconds. Identifies forms, buttons, and API endpoints.
|
|
78
|
+
|
|
79
|
+
### Step 2: Generate
|
|
80
|
+
```bash
|
|
81
|
+
wmcp-annotate generate https://your-site.com --format react
|
|
82
|
+
```
|
|
83
|
+
AI generates WebMCP tool definitions with ready-to-use code.
|
|
84
|
+
|
|
85
|
+
### Step 3: Integrate
|
|
86
|
+
Copy the generated code into your app. Done.
|
|
87
|
+
|
|
88
|
+
### Step 4: Validate
|
|
89
|
+
```bash
|
|
90
|
+
wmcp-annotate validate https://your-site.com --ci
|
|
91
|
+
```
|
|
92
|
+
Add to CI. Never ship broken WebMCP again.
|
|
93
|
+
|
|
94
|
+
---
|
|
95
|
+
|
|
96
|
+
## Pricing Section
|
|
97
|
+
|
|
98
|
+
### Open Source - Free
|
|
99
|
+
- Scan, generate, validate commands
|
|
100
|
+
- 10 AI suggestions/day
|
|
101
|
+
- Community support
|
|
102
|
+
- [Get Started]
|
|
103
|
+
|
|
104
|
+
### Pro - $49/month
|
|
105
|
+
- Unlimited AI suggestions
|
|
106
|
+
- Inject command (auto-add to codebase)
|
|
107
|
+
- CI integration
|
|
108
|
+
- Email support
|
|
109
|
+
- [Start Free Trial]
|
|
110
|
+
|
|
111
|
+
### Enterprise - $499/month
|
|
112
|
+
- Team workspaces
|
|
113
|
+
- SSO/SAML
|
|
114
|
+
- Audit logs
|
|
115
|
+
- Priority support
|
|
116
|
+
- Custom integrations
|
|
117
|
+
- [Contact Sales]
|
|
118
|
+
|
|
119
|
+
---
|
|
120
|
+
|
|
121
|
+
## Social Proof (placeholder)
|
|
122
|
+
|
|
123
|
+
> "We made our entire dashboard AI-agent ready in an afternoon. Game changer."
|
|
124
|
+
> — CTO, Series B SaaS
|
|
125
|
+
|
|
126
|
+
> "The CI integration caught 3 WebMCP regressions in our first week."
|
|
127
|
+
> — Staff Engineer, Fortune 500
|
|
128
|
+
|
|
129
|
+
---
|
|
130
|
+
|
|
131
|
+
## FAQ
|
|
132
|
+
|
|
133
|
+
**What is WebMCP?**
|
|
134
|
+
WebMCP (Web Model Context Protocol) is a W3C standard that lets websites declare structured tools for AI agents. Shipped in Chrome 146, February 2026.
|
|
135
|
+
|
|
136
|
+
**Do I need Chrome 146?**
|
|
137
|
+
For validation, yes. For scanning and generating, any modern browser works.
|
|
138
|
+
|
|
139
|
+
**How does AI analysis work?**
|
|
140
|
+
We use Claude to analyze your page elements and generate meaningful tool definitions. Your data is not stored.
|
|
141
|
+
|
|
142
|
+
**Can I self-host?**
|
|
143
|
+
Yes! The CLI is open source. Pro/Enterprise features require a license key.
|
|
144
|
+
|
|
145
|
+
---
|
|
146
|
+
|
|
147
|
+
## Final CTA
|
|
148
|
+
|
|
149
|
+
### Ready to make your site AI-ready?
|
|
150
|
+
|
|
151
|
+
```bash
|
|
152
|
+
npm install -g wmcp-annotate
|
|
153
|
+
wmcp-annotate scan https://your-site.com
|
|
154
|
+
```
|
|
155
|
+
|
|
156
|
+
[Get Started Free] [Book a Demo]
|
|
157
|
+
|
|
158
|
+
---
|
|
159
|
+
|
|
160
|
+
## Footer
|
|
161
|
+
|
|
162
|
+
Built by Avatar Consulting
|
|
163
|
+
Enterprise AI Transformation for Financial Services
|
|
164
|
+
|
|
165
|
+
[GitHub] [Documentation] [Blog] [Contact]
|
|
@@ -0,0 +1,192 @@
|
|
|
1
|
+
# Social Media Posts - wmcp-annotate
|
|
2
|
+
|
|
3
|
+
## LinkedIn Post
|
|
4
|
+
|
|
5
|
+
---
|
|
6
|
+
|
|
7
|
+
**Google just changed how AI agents interact with websites. Is your site ready?**
|
|
8
|
+
|
|
9
|
+
Two weeks ago, Google and Microsoft shipped WebMCP in Chrome 146—a W3C standard that lets websites declare structured tools for AI agents.
|
|
10
|
+
|
|
11
|
+
Instead of AI scraping your DOM or analyzing screenshots (expensive, fragile), WebMCP lets you tell agents exactly what your site can do.
|
|
12
|
+
|
|
13
|
+
The problem? 99% of sites don't have these annotations yet.
|
|
14
|
+
|
|
15
|
+
So we built wmcp-annotate.
|
|
16
|
+
|
|
17
|
+
One command:
|
|
18
|
+
```
|
|
19
|
+
npx wmcp-annotate generate https://your-site.com
|
|
20
|
+
```
|
|
21
|
+
|
|
22
|
+
It scans your site, uses AI to suggest tool definitions, and generates ready-to-use code.
|
|
23
|
+
|
|
24
|
+
Think of it as "making your website API-first for AI."
|
|
25
|
+
|
|
26
|
+
We're launching today. Open source core, Pro tier for teams.
|
|
27
|
+
|
|
28
|
+
→ GitHub: [link]
|
|
29
|
+
→ Try it: npm install -g wmcp-annotate
|
|
30
|
+
|
|
31
|
+
If you're building AI agents or want your SaaS to be agent-friendly, this is the fastest path.
|
|
32
|
+
|
|
33
|
+
Questions? Drop them below.
|
|
34
|
+
|
|
35
|
+
#AI #WebDevelopment #Automation #AIAgents #WebMCP
|
|
36
|
+
|
|
37
|
+
---
|
|
38
|
+
|
|
39
|
+
## Twitter/X Thread
|
|
40
|
+
|
|
41
|
+
---
|
|
42
|
+
|
|
43
|
+
🧵 Google just shipped the biggest change to web automation since Selenium.
|
|
44
|
+
|
|
45
|
+
WebMCP is now in Chrome 146.
|
|
46
|
+
|
|
47
|
+
Here's what it means and how to get your site ready in 5 minutes:
|
|
48
|
+
|
|
49
|
+
(1/10)
|
|
50
|
+
|
|
51
|
+
---
|
|
52
|
+
|
|
53
|
+
WebMCP lets websites declare "tools" that AI agents can call directly.
|
|
54
|
+
|
|
55
|
+
No more:
|
|
56
|
+
- DOM scraping
|
|
57
|
+
- Screenshot analysis (2000+ tokens each!)
|
|
58
|
+
- Broken automation when CSS changes
|
|
59
|
+
|
|
60
|
+
(2/10)
|
|
61
|
+
|
|
62
|
+
---
|
|
63
|
+
|
|
64
|
+
Instead, your site says:
|
|
65
|
+
|
|
66
|
+
"Here's a searchProducts tool. It takes a query string. Here's how to call it."
|
|
67
|
+
|
|
68
|
+
The AI agent calls the tool. Gets structured data back. Done.
|
|
69
|
+
|
|
70
|
+
(3/10)
|
|
71
|
+
|
|
72
|
+
---
|
|
73
|
+
|
|
74
|
+
This is huge for:
|
|
75
|
+
- AI assistants browsing for users
|
|
76
|
+
- Automation workflows
|
|
77
|
+
- Agent-to-agent commerce
|
|
78
|
+
- Accessibility
|
|
79
|
+
|
|
80
|
+
(4/10)
|
|
81
|
+
|
|
82
|
+
---
|
|
83
|
+
|
|
84
|
+
The catch?
|
|
85
|
+
|
|
86
|
+
99% of websites don't have WebMCP annotations.
|
|
87
|
+
|
|
88
|
+
Someone needs to add them.
|
|
89
|
+
|
|
90
|
+
(5/10)
|
|
91
|
+
|
|
92
|
+
---
|
|
93
|
+
|
|
94
|
+
Enter: wmcp-annotate
|
|
95
|
+
|
|
96
|
+
We built a CLI that:
|
|
97
|
+
1. Scans your site
|
|
98
|
+
2. Identifies forms/buttons/APIs
|
|
99
|
+
3. Uses AI to generate tool definitions
|
|
100
|
+
4. Outputs ready-to-use code
|
|
101
|
+
|
|
102
|
+
(6/10)
|
|
103
|
+
|
|
104
|
+
---
|
|
105
|
+
|
|
106
|
+
```bash
|
|
107
|
+
npx wmcp-annotate generate https://your-site.com
|
|
108
|
+
```
|
|
109
|
+
|
|
110
|
+
That's it.
|
|
111
|
+
|
|
112
|
+
You get JavaScript you can copy-paste into your app.
|
|
113
|
+
|
|
114
|
+
(7/10)
|
|
115
|
+
|
|
116
|
+
---
|
|
117
|
+
|
|
118
|
+
We also added:
|
|
119
|
+
|
|
120
|
+
- TypeScript support
|
|
121
|
+
- React/Vue hooks
|
|
122
|
+
- CI validation
|
|
123
|
+
- Compliance reports
|
|
124
|
+
|
|
125
|
+
(8/10)
|
|
126
|
+
|
|
127
|
+
---
|
|
128
|
+
|
|
129
|
+
Open source core. Free to use.
|
|
130
|
+
|
|
131
|
+
Pro tier ($49/mo) for unlimited AI + CI integration.
|
|
132
|
+
|
|
133
|
+
Enterprise for teams ($499/mo).
|
|
134
|
+
|
|
135
|
+
(9/10)
|
|
136
|
+
|
|
137
|
+
---
|
|
138
|
+
|
|
139
|
+
Try it now:
|
|
140
|
+
|
|
141
|
+
npm install -g wmcp-annotate
|
|
142
|
+
|
|
143
|
+
GitHub: [link]
|
|
144
|
+
Docs: wmcp-annotate.dev
|
|
145
|
+
|
|
146
|
+
If you're building AI agents, this is essential.
|
|
147
|
+
|
|
148
|
+
(10/10)
|
|
149
|
+
|
|
150
|
+
---
|
|
151
|
+
|
|
152
|
+
## Hacker News Post
|
|
153
|
+
|
|
154
|
+
---
|
|
155
|
+
|
|
156
|
+
**Title:** Show HN: wmcp-annotate – Make any website AI-agent ready with WebMCP
|
|
157
|
+
|
|
158
|
+
**Text:**
|
|
159
|
+
|
|
160
|
+
Hey HN,
|
|
161
|
+
|
|
162
|
+
Two weeks ago, Google and Microsoft shipped WebMCP in Chrome 146 (https://webmcp.link). It's a W3C standard that lets websites declare structured tools for AI agents—no more DOM scraping or screenshot parsing.
|
|
163
|
+
|
|
164
|
+
We built wmcp-annotate to help developers adopt it quickly.
|
|
165
|
+
|
|
166
|
+
**What it does:**
|
|
167
|
+
- Scans any website and identifies forms, buttons, API calls
|
|
168
|
+
- Uses Claude to generate meaningful tool definitions
|
|
169
|
+
- Outputs ready-to-use JS/TS/React code
|
|
170
|
+
- Validates existing implementations
|
|
171
|
+
|
|
172
|
+
**Usage:**
|
|
173
|
+
```
|
|
174
|
+
npm install -g wmcp-annotate
|
|
175
|
+
wmcp-annotate generate https://your-site.com --format typescript
|
|
176
|
+
```
|
|
177
|
+
|
|
178
|
+
**Why this matters:**
|
|
179
|
+
|
|
180
|
+
AI agents currently waste ~2000 tokens per screenshot to understand a page. With WebMCP, it's ~100 tokens to call a declared tool. That's 95% cost reduction.
|
|
181
|
+
|
|
182
|
+
Plus, the automation is stable—no more broken scripts when CSS changes.
|
|
183
|
+
|
|
184
|
+
**Business model:**
|
|
185
|
+
|
|
186
|
+
Open source CLI (MIT). Pro tier ($49/mo) for unlimited AI analysis and CI integration. Enterprise ($499/mo) for teams.
|
|
187
|
+
|
|
188
|
+
We're a small consulting firm (Avatar Consulting) that does AI transformation for banks. This came out of client work—everyone wanted their internal tools to be agent-accessible.
|
|
189
|
+
|
|
190
|
+
Would love feedback on the tool and the approach. What's missing?
|
|
191
|
+
|
|
192
|
+
GitHub: [link]
|
package/package.json
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
1
|
+
{
|
|
2
|
+
"name": "wmcp-annotate",
|
|
3
|
+
"version": "1.0.0",
|
|
4
|
+
"description": "Make any website AI-agent ready with WebMCP annotations",
|
|
5
|
+
"main": "dist/index.js",
|
|
6
|
+
"bin": {
|
|
7
|
+
"wmcp-annotate": "dist/index.js"
|
|
8
|
+
},
|
|
9
|
+
"type": "module",
|
|
10
|
+
"scripts": {
|
|
11
|
+
"build": "tsc",
|
|
12
|
+
"dev": "tsx src/index.ts",
|
|
13
|
+
"test": "vitest",
|
|
14
|
+
"test:coverage": "vitest --coverage",
|
|
15
|
+
"lint": "eslint src --ext .ts",
|
|
16
|
+
"prepublishOnly": "npm run build"
|
|
17
|
+
},
|
|
18
|
+
"keywords": [
|
|
19
|
+
"webmcp",
|
|
20
|
+
"ai",
|
|
21
|
+
"agents",
|
|
22
|
+
"automation",
|
|
23
|
+
"browser",
|
|
24
|
+
"cli",
|
|
25
|
+
"w3c",
|
|
26
|
+
"mcp",
|
|
27
|
+
"model-context-protocol"
|
|
28
|
+
],
|
|
29
|
+
"author": "Avatar Consulting <hello@avatarconsulting.com>",
|
|
30
|
+
"license": "MIT",
|
|
31
|
+
"repository": {
|
|
32
|
+
"type": "git",
|
|
33
|
+
"url": "https://github.com/avatarconsulting/wmcp-annotate"
|
|
34
|
+
},
|
|
35
|
+
"homepage": "https://wmcp-annotate.dev",
|
|
36
|
+
"bugs": {
|
|
37
|
+
"url": "https://github.com/avatarconsulting/wmcp-annotate/issues"
|
|
38
|
+
},
|
|
39
|
+
"engines": {
|
|
40
|
+
"node": ">=20.0.0"
|
|
41
|
+
},
|
|
42
|
+
"dependencies": {
|
|
43
|
+
"@anthropic-ai/sdk": "^0.30.0",
|
|
44
|
+
"ajv": "^8.17.0",
|
|
45
|
+
"chalk": "^5.3.0",
|
|
46
|
+
"commander": "^12.1.0",
|
|
47
|
+
"handlebars": "^4.7.8",
|
|
48
|
+
"ora": "^8.0.0",
|
|
49
|
+
"playwright": "^1.48.0"
|
|
50
|
+
},
|
|
51
|
+
"devDependencies": {
|
|
52
|
+
"@types/node": "^22.0.0",
|
|
53
|
+
"eslint": "^9.0.0",
|
|
54
|
+
"tsx": "^4.19.0",
|
|
55
|
+
"typescript": "^5.6.0",
|
|
56
|
+
"vitest": "^2.1.0"
|
|
57
|
+
}
|
|
58
|
+
}
|
|
@@ -0,0 +1,33 @@
|
|
|
1
|
+
#!/bin/bash
|
|
2
|
+
# wmcp-annotate publish script
|
|
3
|
+
# Run this after creating the GitHub repo
|
|
4
|
+
|
|
5
|
+
set -e
|
|
6
|
+
|
|
7
|
+
echo "🚀 wmcp-annotate publish script"
|
|
8
|
+
echo "================================"
|
|
9
|
+
|
|
10
|
+
# Check npm login
|
|
11
|
+
echo "Checking npm login..."
|
|
12
|
+
if ! npm whoami &> /dev/null; then
|
|
13
|
+
echo "❌ Not logged into npm. Running npm login..."
|
|
14
|
+
npm login
|
|
15
|
+
fi
|
|
16
|
+
|
|
17
|
+
# Build
|
|
18
|
+
echo "Building..."
|
|
19
|
+
npm run build
|
|
20
|
+
|
|
21
|
+
# Publish to npm
|
|
22
|
+
echo "Publishing to npm..."
|
|
23
|
+
npm publish --access public
|
|
24
|
+
|
|
25
|
+
echo ""
|
|
26
|
+
echo "✅ Published to npm!"
|
|
27
|
+
echo ""
|
|
28
|
+
echo "Next steps:"
|
|
29
|
+
echo "1. Create GitHub repo: gh repo create avatarconsulting/wmcp-annotate --public"
|
|
30
|
+
echo "2. Push code: git remote add origin git@github.com:avatarconsulting/wmcp-annotate.git && git push -u origin master"
|
|
31
|
+
echo "3. Deploy landing page to Vercel/Netlify"
|
|
32
|
+
echo "4. Post to social media (see LAUNCH.md)"
|
|
33
|
+
echo ""
|